Repeated Administration of 2-Hydroxypropyl-β-Cyclodextrin (HPβCD) Attenuates the Chronic Inflammatory Response to Experimental Stroke

Globally, more than 67 million people are living with the effects of ischemic stroke. Importantly, many stroke survivors develop a chronic inflammatory response that may contribute to cognitive impairment, a common and debilitating sequela of stroke that is insufficiently studied and currently untreatable. 2-Hydroxypropyl-β-cyclodextrin (HPβCD) is an FDA-approved cyclic oligosaccharide that can solubilize and entrap lipophilic substances. The goal of the present study was to determine whether the repeated administration of HPβCD curtails the chronic inflammatory response to stroke by reducing lipid accumulation within stroke infarcts in a distal middle cerebral artery occlusion mouse model of stroke. To achieve this goal, we subcutaneously injected young adult and aged male mice with vehicle or HPβCD 3 times per week, with treatment beginning 1 week after stroke. We evaluated mice at 7 weeks following stroke using immunostaining, RNA sequencing, lipidomic, and behavioral analyses. Chronic stroke infarct and peri-infarct regions of HPβCD-treated mice were characterized by an upregulation of genes involved in lipid metabolism and a downregulation of genes involved in innate and adaptive immunity, reactive astrogliosis, and chemotaxis. Correspondingly, HPβCD reduced the accumulation of lipid droplets, T lymphocytes, B lymphocytes, and plasma cells in stroke infarcts. Repeated administration of HPβCD also preserved NeuN immunoreactivity in the striatum and thalamus and c-Fos immunoreactivity in hippocampal regions. Additionally, HPβCD improved recovery through the protection of hippocampal-dependent spatial working memory and reduction of impulsivity. These results indicate that systemic HPβCD treatment following stroke attenuates chronic inflammation and secondary neurodegeneration and prevents poststroke cognitive decline. SIGNIFICANCE STATEMENT Dementia is a common and debilitating sequela of stroke. Currently, there are no available treatments for poststroke dementia. Our study shows that lipid metabolism is disrupted in chronic stroke infarcts, which causes an accumulation of uncleared lipid debris and correlates with a chronic inflammatory response. To our knowledge, these substantial changes in lipid homeostasis have not been previously recognized or investigated in the context of ischemic stroke. We also provide a proof of principle that solubilizing and entrapping lipophilic substances using HPβCD could be an effective strategy for treating chronic inflammation after stroke and other CNS injuries. We propose that using HPβCD for the prevention of poststroke dementia could improve recovery and increase long-term quality of life in stroke sufferers.


Introduction
Ischemic stroke is a leading cause of death and disability worldwide (Virani et al., 2020). Of stroke survivors, approximately one-third develop a delayed and progressive form of cognitive decline (Black, 2011); however, there are currently no targeted pharmacological interventions to prevent chronic neurodegeneration after stroke. With an aging global population, there is an increasing need for novel and effective treatments that promote recovery after stroke.
We previously reported, in a mouse model of poststroke dementia, that delayed cognitive impairment is caused by a chronic inflammatory response to stroke that is mediated in part by B lymphocytes (Doyle et al., 2015(Doyle et al., , 2012. In addition to the progressive infiltration of B lymphocytes, T lymphocytes, and IgA 1 plasma cells into the infarct in the weeks following stroke, the chronic inflammatory response is also characterized by the production of neurotoxic molecules, such as antibodies, cytokines, and degradative enzymes (Zbesko et al., 2021. These neurotoxic molecules permeate the glial scar and promote chronic inflammation and secondary neurodegeneration in the surrounding parenchyma . Notably, the chronic inflammatory response to stroke has a similar cellular and molecular profile to atherosclerosis. Chronic stroke infarcts contain foamy macrophages, lipid droplets, and intracellular and extracellular cholesterol crystals . In atherosclerosis, these distinguishing characteristics are caused by overwhelmed lipid processing systems within macrophages, resulting in the recruitment of adaptive immune cells and the production of proinflammatory cytokines and degradative enzymes. Lipids are principal structural components of the myelin sheath and are major constituents of the human brain (Vance, 2012). Therefore, it is likely that foamy macrophages and cholesterol crystals form in infarcts following ischemic stroke because lipid debris derived from the breakdown of myelin and other cell membranes overwhelms the processing capacity of infiltrating macrophages and resident microglia (Cantuti-Castelvetri et al., 2018). 2-Hydroxypropyl-b -cyclodextrin (HPb CD) is an FDAapproved compound that entraps lipids and promotes liver X receptor-mediated transcriptional reprogramming in macrophages to improve cholesterol efflux and incite anti-inflammatory mechanisms (Zimmer et al., 2016). Importantly, HPb CD reduces lipid levels in the lesion following spinal cord injury (SCI) (Mar et al., 2016) and prevents lipid overload within phagocytic cells in atherosclerosis and Niemann-Pick disease type C (Taylor et al., 2012;Matsuo et al., 2013;Zimmer et al., 2016). Thus, administration of HPb CD is a potential multipronged approach for mitigating lipid accumulation in phagocytic cells following ischemic stroke.
Therefore, the objective of this study was to investigate the efficacy of HPb CD as a prospective treatment for chronic inflammation and delayed cognitive impairment following ischemic stroke. The aims of the study were twofold: (1) to characterize the lipidome of chronic stroke infarcts; and (2) to determine whether lipid complexation and macrophage reprogramming within infarcts, via the repeated systemic administration of HPb CD, attenuates chronic inflammation and secondary neurodegeneration following stroke. To accomplish these aims, we used the distal middle cerebral artery (MCA) occlusion 1 hypoxia (DH) mouse model of ischemic stroke in conjunction with immunohistochemistry, RNA-Seq, lipidomic, and behavioral analyses. Our resulting data indicate that, coincident with the substantial accumulation of infiltrating immune cells (Doyle et al., 2015;Zbesko et al., 2021), chronic stroke infarcts amass lipids, including sphingomyelins, cholesterol esters, and sulfatides. We also demonstrate that repeated administration of HPb CD in young adult and aged male mice attenuates immune cell and lipid accumulation within chronic stroke infarcts and improves recovery at transcriptional and functional levels. Therefore, HPb CD has the potential to improve stroke recovery and prevent poststroke dementia in humans.

Materials and Methods
Experimental design and statistical analysis. The study was designed in a manner consistent with Stroke Therapy Academic Industry Roundtable and RIGOR guidelines (Stroke Therapy Academic Industry Roundtable, 1999;Landis et al., 2012;Lapchak et al., 2013). Sample sizes for each analysis were determined using power analyses based on expected variances and group differences. Allocation of mice into treatment groups was performed randomly, and behavior experiments were performed with the experimenter blind to treatment group. Statistical analyses were performed with Prism 6.0 (GraphPad). Data are mean 6 SEM. Group sizes, statistical tests, and p values for each experiment are reported in Table 1. For complete blood count (CBC) and lipidomic analyses, normality was assessed using the Kolmogorov-Smirnov test. For all other analyses, data from each individual mouse are presented so the distribution within each group can be visually evaluated; however, normality was not formally assessed. For gene set enrichment analysis (GSEA), the normalized enrichment score is reported. The normalized enrichment score accounts for differences in gene set size and in correlations between gene sets. The normalized enrichment score is based on all dataset permutations to correct for multiple hypothesis testing.
Mice. Young adult (3-to 4-month-old) and aged (17-to 18-monthold) WT male C57BL/6J mice (stock #000664) were purchased from The Jackson Laboratory. Mice were housed in a temperature-controlled suite under a 12 h light-dark regimen, with food and water available ad libitum. All procedures met National Institutes of Health guidelines and were approved by the University of Arizona Institutional Animal Care and Use Committee. At the termination of each experiment, mice were killed by isoflurane anesthesia (JD Medical), exsanguination, and subsequent intracardial perfusion with 0.9% saline. Whole brains or individual brain regions were then removed and placed in RNAlater (Invitrogen, catalog #AM7020) for RNA sequencing analysis, flash frozen in liquid nitrogen for lipidomic analysis, or placed in a 4% PFA solution for 24 h before being transferred into a 30% sucrose solution for immunostaining analysis. Whole blood collected at the time of death was either placed directly in heparin-coated tubes for CBC analysis or extracted with sodium citrate to separate plasma via centrifugation. Stroke surgeries. Stroke was induced in mice using the DH model. The DH stroke model generates a sizable infarct (24% of the ipsilateral hemisphere centered on the somatosensory cortex), has low variability, and has exceptional long-term survivability (Doyle et al., 2012;Nguyen et al., 2016). To induce stroke, we anesthetized animals by isoflurane inhalation and kept them at 37°C throughout the surgical procedure. For all experiments, we injected mice subcutaneously with a single dose of buprenorphine hydrochloride (0.1 mg/kg; Henry Schein Medical) and a single dose of cefazolin antibiotic (25 mg/kg; Sigma-Aldrich) dissolved in sterile saline. Following preoperative preparation, an incision was made to expose the right temporoparietal skull between the orbit and the ear. Under an operating microscope, a small hole was made with a high-speed microdrill through the outer surface of the semitranslucent skull over the visually identified MCA at the level of the parietal cerebral artery. Permanent occlusion of the MCA was performed by electrocoagulation with a small-vessel cauterizer (Bovie Medical). Surgical wounds were closed using Surgi-lock 2oc tissue adhesive (Meridian Animal Health). We then immediately transferred mice to a hypoxia chamber (Coy Laboratory Products) containing 9% oxygen and 91% nitrogen for 45 min. Sustained-release buprenorphine (Bup-SR, 1 mg/kg s.c.; ZooPharm) was administered 24 h after surgery as postoperative analgesia.
MRI. Infarct volume was assessed by MRI 24 h after stroke using a Bruker Biospec 70/20 7.0T scanner with ParaVision-360.2.0 software and a 4-channel phase array mouse coil (Bruker Biospin). Mice were placed in a cradle equipped with a stereotaxic frame, an integrated heating system to maintain body temperature at 37 6 1°C, and a pressure probe to monitor respiration. During MRI acquisition, anesthesia was maintained by inhalation of 1.5%-3% isoflurane. High-resolution structural images were acquired using a T 2 -weighted RARE multiecho Bruker pulse sequence with the following parameters: TR = 2000 ms; flip angle = 180°; RARE factor = 12; matrix size = 128 Â 128; averages = 2; FOV = 1.92 cm Â 1.92 cm; slice thickness = 0.8 mm; number of slices = 15; acquisition time = 8 min 53 s. Infarcts and hemispheric cross sections were manually delineated on T 2 -weighted images using Mango version 4.1.
Drug treatment. We injected HPb CD or vehicle subcutaneously 3 times per week, beginning 7 d after stroke surgery, for 6 weeks until death at 7 weeks after stroke. Mice in the treatment group received HPb CD powder (Sigma-Aldrich, catalog #H-107) dissolved in sterile PBS at a dose of 4 g HPb CD/kg body weight. Mice in the vehicle control group received 300 ml sterile PBS.
Sections were imaged using a Keyence BZ-X700 digital microscope with phase contrast, light, and fluorescence capabilities.
Oil Red O (ORO) staining. Frozen tissue sections were mounted on slides and allowed to dry. Mounted sections were dehydrated in 100% propylene glycol for 5 min and then stained with ORO for 10 min (Abcam, catalog #ab150678). The sections were then differentiated in 85% propylene glycol for 3 min and rinsed with distilled water 3 times. Coverslips were applied to all slides using an aqueous mounting medium (Vector Laboratories, catalog #H-1400). Stained sections were imaged with a Keyence BZ-X700 digital microscope.
Alcian blue staining. Frozen tissue sections were mounted on slides and allowed to dry. Slides were immersed in eosin (VWR, catalog #95057-848) and then rinsed 8 times in PBS. Slides were incubated in 3% acetic acid, pH 2.5, for 3 min and then placed in a solution of 1% w/v Alcian blue in 3% acetic acid, pH 2.5, for 30 min (EMD Millipore, catalog #TMS-010-C). Slides were then rinsed in 3% acetic acid and allowed to dry overnight. Slides were cleared in xylene and preserved with Entellan mounting medium (EMD Millipore, catalog #14802) and coverslips. Stained sections were imaged with a Keyence BZ-X700 digital microscope.
Quantification of staining. To quantify the amount of positive staining for CD3« , B220, ORO, and Alcian blue, we used Fiji (Schindelin et al., 2012). The Color Deconvolution function was used to separate color channels in images of Alcian blue staining. We then converted all stitched images to 8-bit and traced the stroke infarct using the Polygon selections tool. We applied a threshold of positive staining to each image and measured the pixel area of positive staining within the selected region. To quantify IgA and CD138, we used the Cell Counter plugin to manually count the number of positive cells within the stroke infarct. To quantify NeuN and c-Fos immunoreactivity in vehicle-and HPb CDtreated mice, we used Fiji to perform global thresholding in conjunction with a watershed algorithm on the ipsilateral side and equivalent region in the contralateral hemisphere. Fields covering the selected regions were analyzed at bregma 0.38 mm and À1.46 mm (NeuN) or À1.82 mm (c-Fos) using a 20Â objective lens. We also measured hippocampal area in images captured with a 2Â objective lens by tracing the hippocampus with the Polygon selections tool in Fiji.
RNA sequencing and data analysis. Infarcts were visually identified and dissected using forceps and a razor blade. Following infarct removal, adjacent peri-infarct regions were dissected. Corresponding regions in the contralateral hemisphere were also dissected. Fresh brain tissue collected from vehicle-and HPb CD-treated mice was immersed in RNAlater (Invitrogen, catalog #AM7020) and delivered to the University of Arizona Genetics Core. Samples were assessed for quality with an Advanced Analytics Fragment Analyzer (High Sensitivity RNA Analysis Kit, catalog #DNF-491/User Guide DNF-491-2014AUG13) and quantity with a Qubit RNA HS Assay Kit (catalog #Q32852). Given satisfactory quality (RNA integrity number .8) and quantity, samples were used for library construction with the TruSeq Stranded mRNA Library Prep Kit from Illumina (catalog #20020595), as well as the KAPA Dual-Indexed Adapter Kit from Roche (catalog #8278555702). Upon completion of library construction, samples were assessed for quality and average fragment size with the Advanced Analytics Fragment Analyzer (High Sensitivity NGS Analysis Kit, catalog #DNF-846/User Guide DNF-486-2014MAR10). Quantity was assessed with an Illumina Universal Adaptor-specific qPCR kit from KAPA Biosystems (KAPA Library Quantification Kit for Illumina NGS, catalog #KK4824/KAPA Library Quantification Technical Guide-AUG2014). Following the final library quality control, samples were equimolar-pooled and clustered for paired-end sequencing on the Illumina NextSeq500 machine to generate 75 bp reads. The sequencing run was performed using Illumina NextSeq500 run chemistry (NextSeq500/550 High Output version 2 Kit 150 cycles, catalog #FC-404-2002). Sequencing data are publicly available at the National Center for Biotechnology Information through Gene Expression Omnibus accession numbers: GSE173544 and GSE173715. For data analysis, the resulting sequences were demultiplexed using bcl2fastq version 2.19 (Illumina) and trimmed of their indexing adaptors using Trimmomatic version 0.32 (Bolger et al., 2014). The trimmed reads were aligned to the GRCm38 reference genome using STAR version 2.5.2b (Dobin et al., 2013). Gene expression was calculated using the htseq-count function of the HTSeq python tool (Anders et al., 2015). Genes were annotated using the BioMart database. Differential expression (DE) analysis of count tables was performed using DESeq2 (Love et al., 2014). GSEA was performed on all significant differentially expressed genes (false discovery rate [FDR]-adjusted p , 0.05) using a database of Gene Ontology (GO) terms for biological processes. Enrichment maps were constructed from GO terms using the Enrichment Map Cytoscape application (Merico et al., 2010). Minimal editing, such as repositioning of nodes and removal of repetitive gene-sets, was performed to optimize the map layout. Pathway analysis was performed on all differentially expressed genes using Ingenuity Pathway Analysis version 01-13 (IPA).
Targeted lipidomic analysis (LIPID MAPS Lipidomics Core). Fresh brain tissue dissected from vehicle-and HPb CD-treated mice was flash frozen and delivered to the LIPID MAPS Lipidomics Core at the University of California, San Diego. Upon arrival, each sample underwent lipid extraction and quality control analyses. The comprehensive sphingolipid panel and cholesterol ester panel were independently performed using established protocols (Quehenberger et al., 2010).
Y-maze spontaneous alternation behavior (SAB) test. Spatial working memory was assessed with the SAB paradigm. Testing occurred in a Yshaped maze consisting of two symmetrical arms and one longer arm arranged at 120°angles (symmetrical arms: 7.5 cm wide, 37.0 cm long, 12.5 cm high; longer arm: 7.5 cm wide, 42.0 cm long, 12.5 cm high). Mice were placed at the end of the longest arm and allowed to freely explore the three arms in a 5 min trial, as described previously (Doyle et al., 2015). The number of arm entries and the number of correct alternations were recorded by an ANY-maze behavioral video tracking software (Stoelting). An entry was recorded when all four limbs of the mouse were within an arm.
Light/dark transition test. The light/dark box arena consisted of a Plexiglas box (40 cm wide, 40 cm long, 35 cm high) divided by a small underpass into two equally sized compartments: a brightly illuminated zone (390 lux) and a covered dark zone (2 lux). Before the test, animals were habituated to the dark for 30 min. Mice were then placed in the dark chamber of the arena and allowed to move freely between the two chambers for 10 min. The time spent in each chamber, the latency to exit the dark chamber, and the total number of transitions were recorded by an ANY-maze behavioral video tracking software (Stoelting).

Results
Lipid composition of chronic stroke infarcts in young adult mice Before evaluating the lipid composition of chronic stroke infarcts, we performed T 2 -weighted MRI on young adult mice 24 h after stroke to visualize the acute area of infarction generated by the DH stroke model. Infarcts were centered on the somatosensory cortex and included a portion of the M1 region of the primary motor cortex (Doyle et al., 2012). Additionally, infarcts did not extend to the hippocampus (Fig. 1A).
We previously reported that, similar to atherogenesis, chronic stroke infarcts accumulate foamy macrophages, cholesterol crystals, and lipid droplets, which results in the upregulation of osteopontin, matrix metalloproteinases, and pro-inflammatory cytokines . To further evaluate the similarities between the pathophysiology of chronic stroke infarcts and atherosclerosis, we evaluated lipid droplet accumulation in chronic stroke infarcts. In the context of atherosclerosis, macrophages frequently become foamy in appearance because of the accumulation of cholesterol esters stored within lipid droplets. In addition to their role as storage containers within foamy macrophages, lipid droplets also contribute directly and indirectly to the pathology of progressing atherosclerotic plaques (Goldberg et al., 2018). To visualize lipid droplets in chronic stroke infarcts, we performed ORO staining on brain sections from young adult mice at 7 weeks after stroke. ORO staining revealed more intracellular lipid droplets in the infarcts compared with the contralateral cortices (Fig. 1B). These findings provide additional evidence that chronic stroke infarcts and atherosclerotic plaques share several similar pathophysiological features.
To assess whether lipid accumulation in chronic stroke infarcts is derived from myelin breakdown, we performed Alcian blue staining on brain sections from young adult mice at 7 weeks after stroke. Alcian blue stains for sulfatide, a major constituent of CNS myelin. Importantly, sulfatide was recently identified as a myelinassociated inhibitor of neurite outgrowth (Winzeler et al., 2011). Chronic stroke infarcts contained an abundance of sulfatides at 7 weeks after stroke (Fig. 1C), indicating that phagocytic cells, including resident microglia and infiltrating macrophages, are strained in their capacity to process and eliminate myelin lipid debris for at least 7 weeks after stroke.
To further evaluate myelin-derived lipids in chronic stroke infarcts, we performed a targeted lipidomic analysis to assess sphingomyelin levels within infarcts compared with contralateral cortices of young adult mice at 7 weeks after stroke. Importantly, sphingomyelins play a central role in the structure of myelin, constituting 4% of myelin lipid content. The accumulation of sphingomyelin is a hallmark of Niemann-Pick disease and leads to changes in the plasma membrane that promote neurodegeneration (Mar et al., 2016). Sphingomyelins are also involved in signal transduction pathways and in the regulation of cholesterol and protein trafficking to myelin (Poitelon et al., 2020). In addition to their role in myelin architecture, sphingomyelins are involved in microglial activation and inflammation (Fitzner et al., 2020;Yang et al., 2020). Chronic stroke infarcts were enriched in sphingomyelins 7 weeks after stroke (Fig. 1D), indicating a pronounced dysregulation of myelin lipid catabolism in chronic stroke infarcts of young adult mice.

Lipid composition of chronic stroke infarcts in aged mice
Stroke remains the leading cause of long-term disability in people over the age of 65 (Virani et al., 2020). Therefore, we used aged (18-month-old) mice to further characterize the lipid composition of chronic stroke infarcts. First, to corroborate our observations in young adult mice, we used ORO to stain for lipid droplets in brain sections collected from aged mice at 7 weeks after stroke. We observed significantly more intracellular lipid droplets in infarcts compared with contralateral cortices ( Fig.  2A). In addition, Alcian blue staining revealed that chronic stroke infarcts in aged mice are also characterized by an abundance of sulfatides compared with contralateral cortices at 7 weeks after stroke (Fig. 2B). These results demonstrate that young adult and aged mice exhibit a comparable accumulation of lipid droplets and sulfatides at 7 weeks after stroke. Together, these results indicate that young adult and aged mice share similar impediments in the clearance of myelin lipid debris.
Cholesterol is esterified before storage in cytoplasmic lipid droplets to prevent free cholesterol-associated cell toxicity (Ghosh et al., 2010). Therefore, to evaluate the composition of cholesterol esters stored in ORO-stained lipid droplets, we performed a second targeted lipidomic analysis. For this analysis, we again compared infarcts at 7 weeks after stroke to equivalent regions of the contralateral cortex. All quantified cholesterol ester species were significantly elevated in infarcts of aged mice at 7 weeks after stroke compared with contralateral cortices (Fig. 2C). A, Infarct area per slice along the anterior-posterior axis determined by MRI 24 h after DH stroke (n = 3; 0.8 mm slice thickness). Representative T 2 -weighted images show infarcts (areas of hyperintense signal) centered on the somatosensory cortex, extending to the corpus callosum. B, Representative 2Â and 60Â images of ORO staining show markedly more lipid droplets in infarcts than in contralateral cortices (n ¼ 4). Scale bar, 50 mm. ****p , 0.0001 (paired t test). C, Representative 2Â and 20Â images of Alcian blue staining show significantly more sulfatides in infarcts than in contralateral cortices (n ! 4). Scale bar, 125 mm. ****p , 0.0001 (unpaired t test). D, Targeted lipidomic analysis revealed that infarcts had significantly higher levels of sphingomyelins than equivalent regions of the contralateral cortex at 7 weeks after stroke. Sphingomyelins are annotated as ,SM. ,fatty acid chain length.:,number of double bonds._,retention time (min). (n = 8; multiple t tests, Holm-Sidak correction for multiple comparisons; *p , 0.05). Data are mean 6 SEM. SM, Sphingomyelin.

Transcriptome of the chronic stroke infarct in young adult mice
To systematically characterize the chronic stroke infarct transcriptome, we conducted bulk RNA sequencing (RNA-Seq) on infarcts collected at 7 weeks after stroke and compared them with equivalent regions of the contralateral cortex (Fig. 3A). As expected, DE analysis revealed marked differences between infarcts and contralateral cortices. The resultant volcano plot illustrates the distribution of 3930 upregulated and 2869 downregulated genes (Fig. 3B). The most significantly upregulated genes, including Lpl, Spp1, Cd36, Mmp2, and Mmp19, and the most highly enriched genes, including Cd5l, Mmp3, Mmp12, and Mmp13, indicate a pronounced disturbance in lipid homeostasis in infarcts at 7 weeks after stroke. Importantly, Zhu et al. (2017) similarly identified Cd5l, Lpl, and Cd36 as highly expressed genes in foamy macrophages after SCI (Zhu et al., 2017).
In addition to alterations in lipid metabolism, the transcriptome also revealed a signature of chronic inflammation at 7 weeks after stroke. For instance, activation of the innate immune response is indicated by the upregulation of genes involved in complement cascade activation, including C1s1, C3ar1, and C3, and in phagocytosis, including Tmem119, Hexb, Nlrp3, Ctsb, and Myd88. Further, upregulated genes included Gfap, Lcn2, and Serpina3n, which have been validated as specific markers of reactive astrogliosis in ischemic stroke and LPSinduced neuroinflammation (Zamanian et al., 2012) (Fig. 3C). Together, these genes signify the activation of innate immune cells, including infiltrating macrophages, resident microglia, and astrocytes.
Additionally, activation of the adaptive immune system is indicated by the upregulation of genes associated with T lymphocytes, including Cd2, Cd3e, Gzma, Gzmb, and Prf1, B lymphocytes, including CD19, Cd79a, and Blnk, and antibody-producing plasma cells, including Sdc1 and Jchain (Fig.  3C). These perturbations in innate and adaptive immunity demonstrate that chronic inflammation persists in infarcts for at least 7 weeks after stroke, as we have shown previously (Doyle et al., 2015;Chung et al., 2018;Zbesko et al., 2021).
In conjunction with the upregulation of genes involved in chronic inflammation and lipid metabolism, the DE analysis revealed a downregulation of genes involved in neuronal structure and function. Specifically, the neurofilament subunits Nefh, Nefm, and Nefl were downregulated, which indicates persistent disruption or absence of neuronal cytoskeletons in infarcts at 7 weeks after stroke (Fig. 3C).
We then used IPA software to define upstream regulators and identify altered biological processes based on DE analysis. Differentially expressed genes were associated with the following upregulated biological processes: (1) atherosclerosis signaling, (2) IL-1 signaling, (3) inflammasome pathway, (4) eicosanoid signaling, (5) B-cell receptor signaling, and (6) phospholipases, along with others. In contrast, differentially expressed genes were associated with the following downregulated biological processes: (1) calcium signaling, (2) glutamate receptor signaling, and (3) synaptic LTP (Fig. 4A,B). These altered biological processes indicate that the stroke infarct transcriptome is characterized by chronic inflammation, dysregulated lipid metabolism, and impaired or absent neuronal function at 7 weeks after stroke. In addition, IPA identified lipid metabolic upstream regulators, such as cholesterol, phospholipids, and LDL, as well as immunologic upstream regulators such as MYD88, IL18, CD3, IL1B, TLR4, and TNF (Fig. 4C).
HPbCD attenuates the chronic inflammatory response to stroke in young adult mice To investigate the role of lipid metabolism in the chronic inflammatory response to stroke, we administered subcutaneous injections of 4 g/kg HPb CD or vehicle control to young adult mice triweekly (Monday, Wednesday, and Friday) for 6 weeks, beginning 1 week after stroke (Fig. 5A). We performed immunostaining on vehicle-and HPb CD-treated brain sections at 7 weeks after stroke. These analyses revealed substantially fewer B220 1 B lymphocytes, CD3« 1 T lymphocytes, and CD138 1 and IgA 1 antibody-producing plasma cells in infarcts of HPb CD-treated mice than in infarcts of vehicle-injected mice. In addition, ORO staining revealed that infarcts of HPb CD-treated mice had fewer lipid droplets than those of vehicle-injected mice (Fig. 5B). These differences indicate that repeated administration of HPb CD following stroke attenuates chronic inflammation and lipid droplet formation in young adult mice. Conversely, the accumulation of sulfatides in infarcts of vehicle-and HPb CD-treated mice was comparable (Fig. 5B), suggesting that HPb CD does not aid in the clearance of sulfatides.
HPbCD attenuates the chronic inflammatory response to stroke in aged mice Stroke prevalence and mortality rates increase with advancing age in both males and females (Virani et al., 2020). Therefore, to address age as a biological variable, we performed immunohistochemistry and bulk RNA-Seq analyses on aged (18-month-old) mice treated with HPb CD using the same treatment regimen as in young adult mice (HPb CD or vehicle control 3 times per week for 6 weeks) (Fig. 6A). We discovered that, consistent with young adult mice, aged mice had substantially fewer B220 1 B lymphocytes, CD3« 1 T lymphocytes, and IgA 1 antibody-producing plasma cells in their infarcts following repeated administration of HPb CD. In addition, ORO staining revealed that infarcts of HPb CD-treated mice had significantly less lipid droplet accumulation than vehicle-injected mice at 7 weeks after stroke (Fig. 6B). These results demonstrate that young adult and aged mice similarly exhibit attenuated lipid droplet and immune cell accumulation in chronic stroke infarcts following repeated administration of HPb CD. However, the accumulation of sulfatides in vehicle-and HPb CD-treated aged mice was comparable (Fig. 6B), reinforcing that HPb CD does not aid in the clearance of sulfatides.
HPbCD does not alter peripheral immune cell populations in the blood or spleens of aged mice after stroke The spleen, a secondary lymphoid organ, is a major reservoir of immune cells and a focal point for the immune response to tissue injury. In response to ischemic stroke, splenocytes enter into systemic circulation and migrate to the brain, exacerbating neurodegeneration (Seifert et al., 2012). To determine the effect of HPb CD on peripheral immune cell populations, we used flow cytometry to quantify splenic cell populations. There were no significant differences in CD19 1 B lymphocytes or GL-7 1 germinal center B lymphocytes (Fig. 7A,B), or in IgG1 and IgA isotypes (Fig. 7C), between the spleens of vehicle-and HPb CDtreated mice at 7 weeks after stroke. The CD4 1 T lymphocyte, CD8 1 T lymphocyte, and Treg populations were also not significantly altered by repeated HPb CD administration (Fig. 7D,E). These results demonstrate that splenic cell populations in HPb CD-treated aged mice are not significantly altered compared with those of vehicle-injected mice.
A CBC analysis on blood collected from vehicle-and HPb CD-treated aged mice at 7 weeks after stroke also revealed no significant differences in the number of circulating white blood cells, neutrophils, lymphocytes, monocytes, eosinophils, or basophils (Fig. 7F). Together, the flow cytometry and CBC analyses indicate that, while there were substantially fewer immune cells in infarcts of HPb CD-treated aged mice than in those of vehicle-injected mice at 7 weeks after stroke (Fig. 6B), repeated systemic HPb CD administration does not significantly alter circulating or splenic immune cell populations in aged mice.

HPbCD promotes lipid metabolism and attenuates inflammation in infarcts of aged mice after stroke
We systematically characterized the transcriptomes of stroke infarcts in vehicle-and HPb CD-treated aged mice by performing bulk RNA-Seq on infarcted brain tissue collected 7 weeks after stroke (Fig. 8A). As shown in the volcano and MA plots, the DE analysis revealed 991 upregulated genes and 1876 downregulated genes (Fig. 8B,C). HPb CD treatment upregulated intracellular cholesterol transporters, lipoproteins, and lipoprotein receptors, such as Npc1, Npc2, Apoe, Trem2, Abca1, Vldlr, and Hdlbp, and downregulated Ffar1, Lrpb1, and Dgkb (Fig. 8D,E).
These alterations in lipid metabolism suggest that HPb CD aids in the restoration of lipid homeostasis by 7 weeks after stroke. In addition, the DE analysis revealed a downregulation of genes involved in chemotaxis, including Cx3cr1, Cxcl5, and Ccr6, and innate and adaptive immune responses, including Mmd2, Mzb1, Btla, Jchain, and Igkv5-48. Correspondingly, there was an upregulation of anti-inflammatory genes, such as Tgfb1, Tgfb2, P2ry2, and Il18bp (Fig. 8D). These results indicate that HPb CD attenuates the chronic inflammatory response to stroke in aged mice.  We also identified alterations in genes involved in the composition of membranes, including Olig1, Pmp22, Mal2, Rftn1, and Plp. These alterations suggest that HPb CD modifies lipid rafts and cell membranes by 7 weeks after stroke. Genes involved in angiogenesis were also upregulated, including Nrp1, Vash2, Hpse, Vegfc, Vegfd, and Ndnf (Fig.  8D). In addition, enrichment maps constructed from GO terms revealed an upregulation of lipid metabolism, angiogenesis, and collagen synthesis pathways and a downregulation of immune response pathways (Fig.  9A). These results indicate that HPb CD treatment promotes lipid metabolism and angiogenesis and dampens chronic inflammation in infarcts of aged mice at 7 weeks after stroke.
Next, we performed genome-wide expression analysis of the RNA-Seq data using GSEA, which revealed significant enrichment of various biological processes in infarcts of HPb CD-treated aged mice, including (1) collagen formation, (2) regulation of angiogenesis, (3) lipid catabolic process, (4) regulation of autophagy, and (5) regulation of apoptotic process. Conversely, GSEA identified biological processes more enriched in infarcts of vehicle-injected mice compared with HPb CDtreated mice, including (1) innate immune response, (2) adaptive immune response, (3) immunoglobulin production, (4) phagocytosis, engulfment, and (5) complement activation, classical pathway (Fig. 9B). These results further indicate that infarcts of vehicleinjected mice are defined by increased innate and adaptive immune responses, whereas those of HPb CD-treated mice are defined by increased angiogenic and lipid metabolic processes.
Transcriptome of the peri-infarct region in aged mice at 7 weeks after stroke Before assessing the impact of repeated HPb CD administration on the peri-infarct region of the brain, we first compared the transcriptome of the peri-infarct region to the equivalent region of the contralateral hemisphere in aged mice at 7 weeks after stroke (Fig. 10A). The peri-infarct region was characterized by 3910 differentially expressed genes (Fig. 10B,C). Of these, 2123 were upregulated, many of which were also upregulated in infarcts at 7 weeks after stroke (Fig. 3C). For instance, the infarct and peri-infarct regions were both defined by the upregulation of degradative enzymes, including Mmp2, Mmp3, and Mmp8, TLRs, including Tlr1, Tlr2, and Tlr4, and scavenger receptors, including Cd36, Cd68, and Cxcl16. Additionally, peri-infarct regions were characterized by specific markers of reactive astrogliosis, including Gfap, Lcn2, and Serpina3n (Fig. 10D).
Similar to the transcriptome of the infarct at 7 weeks after stroke, the transcriptome of the peri-infarct region was primarily defined by genes involved in chronic inflammation and dysregulated lipid metabolism. The upregulation of Lpl, Lrp2, Ch25h, Apoe, Trem2, and Abca1 signifies a divergence from lipid homeostasis. Moreover, Ldlr was downregulated, which implies impaired lipoprotein and lipid metabolism (Go and Mani, 2012). In addition, genes involved in complement cascade activation (including C1s1, C3ar1, and C3) and phagocytosis (including Tmem119, Hexb, Nlrp3, and Mmd2) were upregulated in peri- Figure 5. HPb CD attenuates the chronic inflammatory response to stroke in young adult mice. A, Experimental design, 3-month-old mice received subcutaneous injections of 4 g/kg HPb CD or vehicle 3 times per week, beginning 1 week after DH stroke. Brains were dissected at 7 weeks after stroke and processed for immunohistochemistry. B, Representative 40Â images of infarcts in vehicle-or HPb CD-treated young adult mice stained for B lymphocytes (B220), T lymphocytes (CD3« ), and antibody-producing plasma cells (CD138, IgA), followed by representative 60Â images of infarcts stained for lipid droplets (ORO) and 20Â images of infarcts stained for sulfatides (Alcian blue). Quantification of images is shown to the right of each photomicrograph (n = 3-11). *p , 0.05; **p , 0.01; ****p , 0.0001; unpaired t tests. Scale bars: ORO, 50 mm; B220, CD3« , CD138, IgA, Alcian blue, 125 mm. Data are mean 6 SEM.
infarct regions. These alterations in innate immunity coincided with corresponding perturbations in adaptive immunity. The upregulation of Cd3d, Cd3e, Cd3g, Blnk, Sdc1, and Jchain suggests that chronic inflammation persists in peri-infarct regions for at least 7 weeks after stroke (Fig. 10D).
Conversely, the DE analysis revealed that 1787 genes were downregulated in peri-infarct regions compared with equivalent regions of the contralateral hemisphere. These genes were primarily involved in mitochondrial respiration and neuronal function. Specifically, genes involved in mitochondrial fission and respiration, including Cyc1, Coa3, Cox5a, and Mff, and the neurofilament subunits Nefh, Nefm, and Nefl, were downregulated (Fig. 10D). Together, these results demonstrate that the transcriptome of the peri-infarct region is characterized by the upregulation of genes involved in inflammation and lipid metabolism and the downregulation of genes involved in mitochondrial respiration and neuronal function.
Next, we performed genome-wide expression analysis of the RNA-Seq data using GSEA. GSEA revealed significant enrichment of various biological processes in peri-infarct regions of aged mice, including (1) macrophage activation, (2) cytokine production, (3) B-cell receptor signaling pathway, (4) lipoprotein metabolism, and (5) phagocytosis. In contrast, GSEA identified biological processes more significantly enriched in equivalent regions of the contralateral hemisphere, including (1) synaptic vesicle localization, (2) mitochondrion organization, (3) neurotransmitter secretion, (4) myelin sheath, and (5) mitochondrial transport (Fig. 11). These results further indicate that peri-infarct regions are defined by chronic inflammation and dysregulated lipid metabolism, whereas equivalent regions of the contralateral hemisphere are defined by intact neuronal function.

HPbCD attenuates inflammation and increases transcripts associated with neuronal function in peri-infarct regions of aged mice after stroke
To evaluate the impact of HPb CD treatment on the transcriptome of the periinfarct region, we performed bulk RNA-Seq on brain tissue collected 7 weeks after stroke. For this analysis, we compared the peri-infarct regions of vehicle-and HPb CD-treated aged mice (Fig. 12A). The peri-infarct region of HPb CDtreated mice was characterized by 5385 differentially expressed genes (Fig. 12B,  C). The DE analysis revealed that specific markers of reactive astrogliosis, including Gfap, Serpina3n, and Stat3, were downregulated following HPb CD treatment. The peri-infarct region of HPb CD-treated mice was also characterized by the downregulation of multiple degradative enzymes, including Mmp2, Mmp8, and Mmp14, chemokine receptors, including Ccr2, Cx3cr1, and Cxcr4, and scavenger receptors, including Cd36, Cd68, and Cxcl16 (Fig. 12D). Figure 6. HPb CD attenuates the chronic inflammatory response to stroke in aged mice. A, Experimental design, 18-month-old mice received subcutaneous injections of 4 g/kg HPb CD or vehicle 3 times per week, beginning 1 week after DH stroke. Brains were dissected at 7 weeks after stroke and processed for immunohistochemistry or RNA-Seq. B, Representative 40Â images of infarcts in vehicleor HPb CD-treated aged mice stained for B lymphocytes (B220), T lymphocytes (CD3« ), and antibody-producing plasma cells (IgA), followed by representative 60Â images of infarcts stained for lipid droplets (ORO) and 20Â images of infarcts stained for sulfatides (Alcian blue). Quantification of images is shown to the right of each photomicrograph (n = 4-10). *p , 0.05; **p , 0.01; ***p , 0.001; unpaired t tests. Scale bars: ORO, 50 mm; B220, CD3« , IgA, Alcian blue, 125 mm. Data are mean 6 SEM.
In addition, the transcriptomic analysis demonstrated a restoration of lipid homeostasis in peri-infarct regions of HPb CDtreated mice compared with vehicle-injected mice, as evidenced by the downregulation of Lipa, Apoe, Lamp1, Npc1, and Npc2 and upregulation of Nceh1 and Ldlr. A reduction in the innate immune response (identified by Tlr4, Csf1, and C2) and in the adaptive immune response (identified by Sdc1, Gsdmd, and Cd1d1) characterized peri-infarct regions of HPb CD-treated Figure 7. HPb CD does not alter peripheral immune cell populations in the blood or spleens of aged mice after stroke. A, B, Splenocytes from HPb CD-and vehicle-injected aged mice were stained with antibodies against CD19 and GL7. Representative flow cytometry plots (A) and quantification of CD19 1 GL7 -B cells and CD19 1 GL7 1 germinal center B cells as a percentage of total lymphocytes (B) showed no significant differences in splenic B-cell populations between vehicle-and HPb CD-treated aged mice at 7 weeks after stroke. C, Quantification of antibody isotypes (IgG1 and IgA) in the spleen revealed no significant differences between vehicle-and HPb CD-treated aged mice at 7 weeks after stroke. D, E, Splenocytes from HPb CD-and vehicleinjected aged mice were stained with antibodies against TCRb , CD4, CD8, and Foxp3. Representative flow cytometry plots of CD4 1 and CD8 1 T cells (D) and quantification of CD4 1 T cells, CD81 T cells, and Tregs as a percentage of total lymphocytes (for CD4 1 and CD8 1 gates) or TCRb 1 CD4 1 T cells (for Treg gate) (E) revealed no significant differences in splenic T-cell populations between vehicle-and HPb CD-treated aged mice at 7 weeks after stroke (n = 9 or 10). *p , 0.05 (unpaired t tests). F, CBC analysis in vehicle-and HPb CD-treated aged mice at 7 weeks after stroke revealed no significant differences in circulating immune cell populations (n = 6). *p , 0.05 (multiple t tests, Holm-Sidak correction for multiple comparisons). Data are mean 6 SEM. Figure 8. HPb CD promotes lipid metabolism and attenuates inflammation in infarcts of aged mice after stroke. A, Schematic of a mouse coronal brain section following DH stroke, with the analyzed region shaded in gray. B, C, Volcano (B) and MA (C) plots constructed from count data show differences in gene expression between infarcts from vehicle-and HPb CD-treated mice at 7 weeks after stroke (FDR-adjusted p , 0.05; FC . |2|). D, Row-scaled heatmap displaying differentially expressed genes associated with lipid metabolism, inflammation, and angiogenesis (FDR-adjusted p , 0.05; FC . |2|). E, Graphs showing rlog-normalized expression of selected genes related to lipid metabolism (n = 4). *p , 0.05; **p , 0.01; ***p , 0.001; unpaired t tests. FC, Fold change. Figure 9. Biological processes enriched in infarcts of HPb CD-treated aged mice at 7 weeks after stroke. A, Enrichment maps constructed from GO terms revealed a significant upregulation of metabolism, angiogenesis, and collagen synthesis pathways and a significant downregulation of immune response pathways. Pathways are shown as circles (nodes) that are connected with lines (edges) if the pathways share genes. Node colors are based on the enrichment score, and edge sizes are based on the number of genes shared by the connected pathways. B, GSEA revealed significant enrichment of various biological processes in infarcts of HPb CD-treated aged mice (red). Conversely, GSEA identified biological processes more significantly enriched in infarcts of vehicle-injected mice compared with HPb CD-treated mice (blue). mice (Fig. 12D); these alterations are consistent with the attenuated chronic inflammation that characterized infarcts of the HPb CD-treated aged mice (Fig. 8D). These results demonstrate that HPb CD treatment after stroke attenuates chronic inflammation and promotes lipid metabolism in peri-infarct regions of aged mice. Additionally, the DE analysis revealed an upregulation of genes involved in myelin sheath structural integrity, such as Ina, Omg, and Nefl, and synaptic plasticity and neuroplasticity, such as Syn1, Bdnf, Nrg1, Negr1, and Nsg1 (Fig. 12D). These results suggest that HPb CD treatment improves the structure and function of neurons in peri-infarct regions of aged mice after stroke.
Next, we performed genome-wide expression analysis of the RNA-Seq data using GSEA, which revealed a significant enrichment of various biological processes in peri-infarct regions of HPb CD-treated aged mice, including (1) synaptic vesicle Figure 10. Transcriptome of the peri-infarct region in aged mice at 7 weeks after stroke. A, Schematic of a mouse coronal brain section following DH stroke, with analyzed regions indicated. B, C, Volcano (B) and MA (C) plots constructed from count data show differences in gene expression between peri-infarct regions and equivalent regions in the contralateral cortex at 7 weeks after stroke (FDR-adjusted p , 0.05; FC . |2|). D, Row-scaled heatmap displaying differentially expressed genes associated with lipid metabolism, inflammation, and mitochondrial function (FDR-adjusted p , 0.05; FC . |2|). FC, Fold change.
HPbCD attenuates neurodegeneration and improves recovery in aged mice after stroke To assess behavioral outcomes following HPb CD treatment, we used the light/dark transition test and the Y-maze SAB test. The light/dark transition test is based on the innate aversion of mice to brightly illuminated areas and on their spontaneous exploratory behavior in response to mild stressors (Bourin and Hascoët, 2003;Takao and Miyakawa, 2006). Our results revealed that, in the light/dark arena, HPb CD-treated mice spent less time in the brightly illuminated compartment than vehicle-injected mice ( Fig. 14A). Additionally, HPb CD-treated mice were slower to emerge from the dark compartment after being habituated to the dark for 30 min before the test (Fig. 14A). These behaviors indicate decreased impulsivity and behavioral disinhibition in HPb CD-treated mice.
We also assessed hippocampal-dependent spatial working memory at 7 weeks after stroke with the Y-maze SAB test (Yamada et al., 1996). We discovered that vehicle-injected mice developed a delayed cognitive deficit, whereas HPb CD-treated mice displayed neurotypical cognitive function comparable to baseline recordings (Fig. 14B). To assess hippocampal edema, we measured the total hippocampal area of the ipsilateral and contralateral hemispheres in aged mice that received vehicle or HPb CD. Hippocampal enlargement 7 weeks after stroke was attenuated in mice that had received HPb CD (Fig. 14C). Through Pearson correlation analysis, we observed a moderate negative correlation (r = À0.5640) between hippocampal edema and hippocampal-dependent spatial working memory outcomes in the Y-maze SAB test (Fig. 14D). We next quantified NeuN immunoreactivity as a biomarker of secondary neurodegeneration in regions of axonal degeneration. Repeated administration of HPb CD led to the preservation of NeuN immunoreactivity in Figure 11. Biological processes enriched in peri-infarct regions of aged mice at 7 weeks after stroke. GSEA revealed significant enrichment of multiple biological processes in peri-infarct regions of aged mice (red). Conversely, GSEA identified biological processes more significantly enriched in equivalent regions of the contralateral hemisphere (blue). the striatum and thalamus (Fig. 14E,F). These results demonstrate that HPb CD attenuates secondary neurodegeneration and improves cognitive function following stroke.
To investigate neuronal activation in hippocampal regions, we next measured the immunoreactivity of c-Fos, a transcription factor thought to primarily reflect NMDA-mediated neuronal activity (Albertini et al., 2018). We observed an induction of c-Fos expression in the dentate gyrus, CA1, and CA3 areas of HPb CD-treated mice (Fig. 14G). Importantly, Fleischmann et al. (2003) demonstrated a critical role for c-Fos in hippocampaldependent spatial learning and memory, as well as in NMDA receptor-dependent LTP formation (Fleischmann et al., 2003). Together, our results show that an improvement in hippocampal-dependent spatial working memory (Fig. 14B) is associated with an induction of c-Fos in hippocampal regions of HPb CDtreated mice (Fig. 14G) at 7 weeks after stroke.
To corroborate these findings, we further analyzed the results from the GSEA performed on peri-infarct regions from vehicleand HPb CD-treated aged mice (Fig. 13). We discovered that peri-infarct regions of HPb CD-treated aged mice were characterized by enriched biological processes that signify improved neuronal and synaptic function, including (1) long-term synaptic potentiation, (2) neurotransmitter transport, (3) synaptic signaling, (4) glutamate receptor activity, (5) membrane depolarization during action potential, and (6) learning or memory (Fig. 14H). These results provide evidence that HPb CD improves neuronal integrity in peri-infarct regions by 7 weeks after stroke.

Discussion
Dysregulated lipid homeostasis has been implicated in chronic neurodegenerative diseases, including Alzheimer's disease, Figure 12. HPb CD attenuates inflammation and increases transcripts associated with neuronal function in peri-infarct regions of aged mice after stroke. A, Schematic of a mouse coronal brain section following DH stroke, with the analyzed region indicated. B, C, Volcano (B) and MA (C) plots constructed from count data show differences in gene expression between peri-infarct regions from vehicle-and HPb CD-treated aged mice at 7 weeks after stroke (FDR-adjusted p , 0.05; FC . |2|). D, Row-scaled heatmap displaying differentially expressed genes associated with lipid metabolism, inflammation, and neuronal function (FDR-adjusted p , 0.05; FC . |2|). FC, Fold change.
Parkinson's disease, and Niemann-Pick disease type C, and in acute neuronal injuries, including ischemic stroke (Vance, 2012). Lipid metabolism is particularly important in the mammalian brain, as lipids constitute 50%-60% of the dry weight (Luchtman and Song, 2013). Notably, the mammalian brain is highly enriched in cholesterol, which primarily exists in two distinct pools within the CNS: in plasma membranes of glial cells and neurons and in myelin sheaths surrounding axons. Because of the intact blood-brain barrier, cholesterol synthesis and lipoprotein transport are regulated independently of the peripheral circulation (Dietschy and Turley, 2001;Vance, 2012). Therefore, we set out to further assess the lipidome of chronic stroke infarcts.
Our previous preclinical studies indicate that there is substantial overlap between the pathologic characteristics of chronic stroke infarcts and the hallmark features of atherosclerotic plaques. Similar to atherosclerotic plaques, chronic stroke infarcts contain foamy macrophages, lipid droplets, and intracellular and extracellular cholesterol crystals . We have also shown that the development of these pathologies coincides with the recruitment and infiltration of adaptive immune cells (Doyle et al., 2015;Zbesko et al., 2021). Importantly, in atherogenesis, the formation of foamy macrophages, caused by dysregulated lipid metabolic processes, leads to the recruitment of adaptive immune cells and the production of cytokines and degradative enzymes. Herein, we postulated that, following stroke, Figure 13. Biological processes enriched in peri-infarct regions of HPb CD-treated aged mice at 7 weeks after stroke. GSEA revealed significant enrichment of multiple biological processes in peri-infarct regions of HPb CD-treated aged mice (red). Conversely, GSEA identified biological processes more significantly enriched in peri-infarct regions of vehicle-injected aged mice (blue). Figure 14. HPb CD attenuates neurodegeneration and improves recovery in aged mice after stroke. A, HPb CD-treated aged mice exhibited less impulsive behavior than vehicle-injected mice in the light/dark transition test at 7 weeks after stroke (n = 13-15). *p , 0.05 (unpaired t tests). B, HPb CD-treated aged mice exhibited intact spatial working memory in the Y-maze SAB test compared with vehicle-injected aged mice at 7 weeks after stroke (n = 13-15). *p , 0.05 (unpaired t tests). C, Quantification of hippocampal edema at 7 weeks after stroke in vehicleand HPb CD-treated mice (n = 7-9). *p , 0.05 (unpaired t tests). D, Pearson correlation between Y-maze SAB and hippocampal edema at 7 weeks after stroke (n = 12; r = À0.5640; p . 0.05). E, F, Quantification of NeuN immunoreactivity in the striatum and thalamus from vehicle-or HPb CD-treated mice (n = 5). *p , 0.05; **p , 0.01; unpaired t tests. G, Representative 20Â images of c-Fos immunoreactivity in hippocampal regions from vehicle-or HPb CD-treated mice. Quantification of images is shown to the right of each photomicrograph (n = 9 or 10). *p , 0.05; ***p , 0.001; unpaired t tests. Scale bar, 125 mm. Data are mean 6 SEM. H, GSEA revealed significant enrichment of biological processes associated with neuronal function and activity in peri-infarct regions of HPb CD-treated aged mice at 7 weeks after stroke. lipids derived from myelin debris and other cell membranes overwhelm the processing capability of phagocytes in the brain, leading to the formation of lipid-laden foam cells, generation of cholesterol crystals, secretion of pro-inflammatory cytokines, and production of degradative enzymes. Further, we hypothesized that this chronic inflammatory response, coupled with concurrent cell death, causes poststroke secondary neurodegeneration and impairs functional recovery.
Using the DH mouse model of stroke, we first characterized the lipidome of chronic stroke infarcts. The lipidome of infarcts at 7 weeks after stroke showed a substantial elevation of lipids, including sphingomyelins, sulfatides, and cholesterol esters, compared with contralateral cortices. We also observed an accumulation of intracellular lipid droplets. It has been shown that the excessive accumulation of cholesterol esters in lipid droplets contributes to foam cell formation (Yu et al., 2013). These findings suggest that the foam cells identified in chronic stroke infarcts contain an abundance of cholesterol and other myelinderived lipids.
Using transcriptomic analyses, we assessed the impact of lipid dyshomeostasis on chronic stroke infarcts and peri-infarct regions. We anticipated disparity between ipsilateral and contralateral regions. Indeed, the transcriptomes of chronic stroke infarcts and peri-infarct regions were characterized by an upregulation of genes involved in inflammation, reactive astrogliosis, and lipid metabolism. Specifically, the upregulation of TLRs, including Tlr1, Tlr2, and Tlr4, scavenger receptors, including Scarb2, Cd68, Msr1, and Cxcl16, and other lipid mediators, including Npc1, Npc2, Lamp1, Ffar1, Pparg, Apoe, Trem2, and Abca1, indicates a pronounced disruption in lipid homeostasis at 7 weeks after stroke. These transcriptomic analyses suggest that the mechanism of foam cell formation in ischemic stroke shares commonalities with the mechanism extensively characterized in atherosclerosis, albeit distinct in that myelin debris, rather than circulating LDL cholesterol, appears to be the primary source of lipids that overwhelm the processing capacity of infiltrating macrophages and resident microglia.
We next assessed the efficacy of HPb CD in counteracting or reversing foam cell formation and immune cell infiltration following stroke, as HPb CD induces liver X receptor target gene expression in macrophages and leads to an increase in cholesterol transporters that further promote cholesterol efflux (Zimmer et al., 2016). Cyclodextrins are cyclic oligosaccharides comprised of glucose monomers. With hydrophobic interiors and hydrophilic exteriors, cyclodextrins can form inclusion complexes with hydrophobic compounds and yield aqueous solubility and stability across tissues. Cyclodextrins are commonly used as carriers and solubilizing agents for steroids, antivirals, and chemotherapies (Rasheed et al., 2008;Gidwani and Vyas, 2015). In addition, HPb CD has also proven efficacious in the regression of atherosclerotic plaques and the prevention of age-related lipofuscin accumulation (Zimmer et al., 2016;Gaspar et al., 2017). Here, we found that repeated administration of HPb CD after stroke significantly attenuated chronic inflammation in young adult and aged mouse models. Specifically, there were substantially fewer intracellular lipid droplets, B lymphocytes, T lymphocytes, and antibody-producing plasma cells in the infarcts of young adult and aged HPb CD-treated mice than in vehicleinjected mice at 7 weeks after stroke.
We discovered that HPb CD is effective in attenuating B-lymphocyte accumulation in chronic stroke infarcts, which is significant because we have shown that B lymphocytes mediate cognitive dysfunction following stroke (Doyle et al., 2015).
Correspondingly, activated B lymphocytes and autoantibodies have been shown to cause neuropathology in models of experimental autoimmune encephalomyelitis, SCI, and transient MCA occlusion (Raine et al., 1999;Ankeny et al., 2006;Ortega et al., 2015). However, the B-lymphocyte response to stroke is heterogeneous and has also been shown to be beneficial for stroke recovery. For example, it consists of both T-lymphocyte-dependent and -independent responses (Weitbrecht et al., 2021;Zbesko et al., 2021). Additionally, B-lymphocyte depletion following stroke reduces stroke-induced hippocampal neurogenesis and impairs functional recovery (Ortega et al., 2020). Future studies are necessary to elucidate the role of infiltrating B lymphocytes in stroke recovery.
Using transcriptomic analyses, we next assessed transcripts in infarct and peri-infarct regions of HPb CD-treated aged mice. We found that systemic administration of HPb CD distinctively altered the transcriptomes of infarct and peri-infarct regions. In infarcts, HPb CD upregulated genes involved in lipid metabolic processes, including Lipa, Npc1, Apoe, and Abca1; in peri-infarct regions, HPb CD had an inverse effect on these genes. These observations suggest that HPb CD has supported the restoration of lipid homeostasis through transcriptional reprogramming of macrophages, consistent with a previously described mechanism of action (Zimmer et al., 2016). In addition, HPb CD influenced expression profiles in infarcts through the upregulation of genes involved in angiogenesis and in peri-infarct regions through the upregulation of genes involved in neuronal and synaptic structure and function. Together, these results indicate that HPb CD exerts multiple restorative effects on affected brain regions after stroke.
We next evaluated the efficacy of HPb CD in enhancing recovery after stroke. We assessed impulsivity and risk-taking behavior using the light/dark transition test. We have previously demonstrated a chronic impact of stroke on the psychological measurement of impulsivity in the light/dark transition test, resembling the uninhibited, risk-taking behaviors exhibited by patients with Alzheimer and frontotemporal dementias (Liscic et al., 2007;Nguyen et al., 2018). HPb CD-treated mice spent less time in the brightly illuminated chamber and were slower to exit the dark chamber. These behaviors suggest that HPb CD-treated mice act less erratically and impulsively than vehicle-injected mice. We also measured hippocampal-dependent spatial working memory using the Y-maze SAB test. We found that vehicleinjected mice developed a delayed cognitive deficit, whereas HPb CD-treated mice displayed neurotypical cognitive function. This improvement in recovery was associated with a reduction in hippocampal edema, a preservation of neurons in the striatum and thalamus, and an induction of c-Fos expression in hippocampal regions (dentate gyrus, CA1, CA3). Correspondingly, peri-infarct regions of HPb CD-treated mice were defined by an enrichment of biological processes signifying improved neuronal and synaptic integrity. Thus, administration of HPb CD reduced stroke-induced neuropathology and improved cognitive function after stroke.
For decades, putative neuroprotectants have failed in clinical stroke trials despite demonstrating efficacy in preclinical studies. In light of these failures, we followed Stroke Therapy Academic Industry Roundtable guidelines to accelerate the translation of HPb CD from bench to bedside (Stroke Therapy Academic Industry Roundtable, 1999;Landis et al., 2012;Lapchak et al., 2013): (1) we performed and analyzed experiments in a randomized and blinded manner; (2) we calculated sample sizes based on a power analysis before the start of the study; (3) we delayed treatment by 1 week to achieve an optimal therapeutic time window; (4) we selected a dosage and a systemic route of administration to parallel preclinical trials using HPb CD in Niemann-Pick disease type C; (5) we chose a model of stroke in which longterm functional outcomes could be assessed; and (6) we used both young adult and aged animals. However, there are additional experiments that need to be conducted. For example, we have not yet tested the efficacy of HPb CD in female mice. Although we anticipate comparable efficacy in females, future studies are needed to address sex as a biological variable. In addition, HPb CD must still be tested in a model of transient occlusion, in a different laboratory, and in a second species.
In conclusion, we have shown that, coincident with the progressive infiltration of adaptive immune cells, chronic stroke infarcts accumulate lipids, including sphingomyelins, sulfatides, and cholesterol esters. To our knowledge, this substantial disruption in lipid homeostasis has not been previously recognized or investigated in the context of stroke. We also discovered that repeated administration of HPb CD following stroke aids in the restoration of lipid homeostasis, attenuates lipid droplet and immune cell accumulation, and improves recovery at transcriptional and functional levels. Therefore, HPb CD warrants further testing to determine whether it can be repurposed for the treatment of ischemic stroke and other CNS injuries.